Disseminated intravascular coagulopathy in nonpromyelocytic acute myeloid leukemia – incidence, clinical and laboratory features and prognostic significance

Introduction: Acute promyelocytic leukemia (APL) has the highest risk for overt disseminated intravascular coagulopathy (DIC), with reported incidence of DIC of up to 90%, as compared to 10-40% in other AML types. The influence of DIC on early death in non-APL AML patients has not been evaluated so far.

Aim: The aim of our study was to analyze the incidence of DIC, its clinical and laboratory characteristics, and the impact on the survival and early death of patients with non-APL AML.

Materials and methods: A total of 176 patients with non-APL AML, diagnosed and treated at the Clinic for Hematology of the Clinical Center of Serbia, between 2015 and 2020, were evaluated retrospectively. The diagnosis of DIC was made on the basis of ISTH (International Society on Thrombosis and Haemostasias) criteria.

Results: The mean age of our patients was 53.8 ± 14.6 years, with 99/176 patients being men (56.2%). DIC was present in 74/176 patients (42.05%), who had a significant prevalence of the hemorrhagic syndrome (p = 0.01). The risk factors for overt DIC were the following: older age (p <0.01), comorbidities (p = 0.01), leukocytosis (p <0.001) and a high level of LDH (p <0.001). The FAB (French, American and British) type of non-APL AML, the cytogenetic risk group, and CD56 (cluster of differentiation) had no influence on overt DIC (p > 0.05). No difference was found in early mortality, outcome, and the survival of non-APL AML patients, with and without DIC (p > 0.05).

Conclusion: Older age at diagnosis, comorbidities, leukocytosis, and high LDH concentrations are found to be adverse risk factors for overt DIC in non-APL AML patients. If treated promptly, with immediate, adequate and intensive use of blood derivates and components, DIC has no negative impact on early mortality, outcome, and survival.

INTRODUCTION

Acute myeloid leukemias (AML) are a heterogenous group of malignant diseases of the blood characterized by clonal expansion of myeloblasts in the bone marrow (≥20%), in peripheral blood, and/or in other tissues [1]. AML is a rare disease and accounts for 1.1% of all malignant diseases. AML is the most frequently occurring type of acute leukemia of adult age and has an annual incidence of 4.3 per 100,000 population, occurring somewhat more frequently in men (m : f = 5.2/100,000 : 3.6/100,000). AML is a disease of the elderly – at diagnosis, 54% of the patients are 65 years old or above, with the median age being 68 – 71 years [2],[3],[4]. Despite modern treatment, survival of AML patients is very short (five-year survival = 24%) [2].

As far as therapy-related AML (t-AML) is concerned, previous chemotherapy, radiotherapy, and immunosuppressive drugs, for the treatment of malignant and autoimmune diseases, can contribute to its development. Occupational exposure, as well as exposure to agents from the environment, which damage DNA (deoxyribonucleic acid), can also contribute to the occurrence of AML. AML can develop as a secondary disease, i.e., it can occur as the result of the evolution of chronic myeloproliferative neoplasms (MPN) or myelodysplastic syndromes (MDS). Also, a genetic predisposition for the development of AML (Fanconi anemia, Down syndrome, Shwachman-Diamond syndrome, congenital neutropenia syndromes) has been confirmed [1],[5]. However, as far as de novo AML is concerned, the etiology of most AMLs remains unknown.

In AML, there is a whole array of different chromosomal alterations, such as the translocation t(15;17)  (PML-RARA), which is characteristic of acute promyelocytic leukemia (APL). There are also a number of genetic mutations, which affect signal pathways (such as FLT3-ITD,  KIT, MLL, KRAS, NRAS), nucleophosmin (NPM1), transcription factors (such as CEBPA, RUNX1, GATA-2), and tumor suppression (TP53, WT1). In AML, there are also epigenetic mutations which lead to the methylation of DNA and the modification of chromatin (such as TET, IDH1, IDH2, MLL) [5],[6].

The contemporary classification of acute leukemias, issued by the World health Organization (WHO) in 2016, as well as the European LeukemiaNet (ELN) recommendations are, in fact, based on the molecular characteristics of AML, since they have both prognostic and therapeutic significance [7],[8], although, in everyday clinical practice, the French, American and British (FAB) classification system, which is based on morphological and immunophenotypical characteristics of leukemia cells is still in use [9],[10].

Clinically, AML develops from “full health” and presents with elevated body temperature, anemia, bleeding and recurrent infections. AML patients often have thrombocytopenia and hemostasis disorders, i.e., coagulopathies, which significantly complicate treatment and contribute to early mortality of these patients [11].

Disseminated intravascular coagulopathy (DIC) is an acquired syndrome characterized by systemic intravascular activation of coagulation, which can lead to multiorgan dysfunction, thrombosis, and/or excessive bleeding [12],[13]. The most common conditions leading to DIC, are the following: sepsis, shock, solid tumors, and malignant diseases of the blood – acute leukemias and non-Hodgkin lymphomas. Under the influence of proinflammatory cytokines, mononuclear and endothelial cells express the tissue factor (TF). Through the contact of TF with coagulation factors in the blood, the coagulation cascade is initiated, which leads to the generation of thrombin and the conversion of fibrinogen into fibrin. At the same time, interaction between thrombocytes and the blood vessel wall contributes to the creation of vascular (or microvascular) thrombi. P-selectin from activated thrombocytes additionally intensifies the expression of TF. The binding of TF, thrombin, and other activated coagulation factors (proteases) to specific protease-activated receptors (PAR), and the binding of fibrin to toll-like receptor 4 (TLR4) on inflammatory cells, affects inflammation through the consequential release of pro-inflammatory cytokines and chemokines, which further modulates coagulation and fibrinolysis [14].

The International Society on Thrombosis and Hemostasis (ISTH) has recommended a scoring system, which is applied in patients with an underlying disorder, known to be linked to the development of DIC, where the following four laboratory parameters are monitored: platelet (thrombocyte) count (Tr), prothrombin time (PT), fibrinogen concentration, and the D-dimer level [12]. Disseminated intravascular coagulopathy is present in as many as 90% of the patients with APL [15], while the frequency of DIC in other types of AML is significantly lower, and ranges from 10% do 40% [16]. There are no studies analyzing the effect of the ISTH DIC on early mortality in patients with AML.

The aim of our study was to collect and analyze data on the following: the incidence of DIC in a group of patients with non-APL AML, the clinical presentation, clinical and laboratory parameters, i.e., the frequency of bleeding, thrombosis, and early death of non-APL AML patients with DIC.

METHODS

A retrospective analysis was performed, involving 176 consecutive patients with non-APL AML, diagnosed and treated at the Clinic for Hematology of the Clinical Center of Serbia, between 2015 and 2020. AML diagnosis was established on the basis of cytomorphological, immunophenotypical, cytogenetic, and molecular characteristics of bone marrow cells or peripheral blood cells, in keeping with the recommendations of the World Health Organization from 2016 [7]. The morphological diagnosis was based on the FAB classification [9]. During immunophenotypization by means of flow cytometry, in addition to the standard monoclonal antibodies [10], the monoclonal antibody for CD56 (cluster of differentiation), which is characteristic of NK (natural killer) cells, was also applied. The cytogenetic risk assessment (favorable, intermediate, unfavorable) was carried out in keeping with the ELN recommendations [8] through the determination of the karyotype – by means of conventional cytogenetics and molecular characteristics – with the use of the polymerase chain reaction (PCR) method. The diagnosis of t-AML was established in patients with a positive personal anamnesis and medical documentation on previous application of chemotherapy, radiotherapy, and immunosuppressive drugs, for treating malignant or autoimmune diseases. When establishing the diagnosis, the significance of existing associated diseases, i.e., comorbidities was determined, on the basis of the HCT-CI (Hematopoietic cell transplantation specific comorbidity index) score [17].

In all patients, the following laboratory parameters were analyzed: hemoglobin-Hb (g/l), white blood cell count (WBC), i.e., leukocyte count – Le (x10⁹ /l), platelet count, i.e., thrombocyte count - Tr (x10⁹ /l), percentage of myeloblasts in peripheral blood, concentration of lactate dehydrogenase – LDH (U/l), PT, activated partial thromboplastin time (aPTT), fibrinogen, and D-dimer. The normal values for LDH were 220 – 460 U/l, while the normal values for hemostasis parameters were: 75 – 120%, for PT; 25 – 35 s, for aPTT; 2 – 4 g/l, for fibrinogen; <0.5 μg/ml, for D-dimer. The diagnosis of DIC was established on the basis of the ISTH criteria: platelet count (x10⁹ /l) : >100x10⁹ /l = 0,75 % = 0, 50 – 75 % = 1, 1 g/l = 0,<1 g/l = 1. The total score of ≥5 indicates overt disseminated intravascular coagulopathy. In all of the patients, it was assessed whether they had clinical signs of hemorrhagic syndrome at the time of diagnosis.

All of the patients received combined induction cytostatic therapy (doxorubicin and cytosine arabinoside, following the ‘3+7’ and ‘2+5’ regimen, depending on the general performance status and the comorbidity index), and then consolidation, with the application of cytosine arabinoside. Parallel to the treatment of AML, patients with overt disseminated intravascular coagulopathy were treated with blood derivatives and blood components, as per the current recommendations. Transfusions of thrombocytes were applied when the thrombocyte count was <50 x10⁹ /l, in patients with manifest bleeding, and when bleeding was absent, in cases when the thrombocyte count was <20 x10⁹ /l. In patients with the hemorrhagic form of DIC and prolonged PT and aPTT, frozen fresh plasma (FFP) was applied, in the dose of 15ml/kg. Patients with severe hypofibrinogenemia (<1g/l), which persisted despite the application of FFP, also received cryoprecipitate [13]. The following were monitored: treatment outcome (living/deceased), early mortality (lethal outcome occurring between the first day of hospitalization and the conclusion of induction treatment, i.e., discharge from hospital), overall survival of all the patients (OS), as well as whether the existence of DIC had any effect on the outcome and OS.

In statistical analysis, the following methods of descriptive statistics were applied: a) for continuous variables – the arithmetic mean and the standard deviation (SD), i.e., the median and the range, and b) for categorical variables – frequency, expressed in absolute values, and percentages. For determining the difference between the two groups, the appropriate statistical tests were applied, namely: the parametric Student’s T-Test for two independent samples, i.e., its non-parametric parallel – the rank sum test (Mann–Whitney U test). For testing the difference in frequency, the chi-square test, i.e., Fisher’s exact probability test were used. For analyzing survival, the Kaplan Meier method, as well as the log-rank test for comparing survival amongst the tested groups, were applied. The values p < 0.05 were believed statistically significant.

RESULTS

The study included 176 patients with non-APL AML, 99 men (56.2%) and 77 women (43.7%) (M : Ž = 1.29), whose average age was 53.8 ± 14.6 years. The demographic, laboratory and clinical characteristics of the patients are presented in Table 1.

Table 1. Multivariate logistic regression models with unmet dental health care needs as an outcome variable

At the time of diagnosis, hemorrhagic syndrome was present in 72/176 patients (40.9%). The analysis of the blood count parameters showed that the patients, on average, had moderate anemia (97.3 ± 18.4 g/l), grade 3 thrombocytopenia (median: 44 x10⁹ /l; range: 1 – 421), and leukocytosis (median: 18.5 x10⁹ /l; range: 0.6 – 473.2), with the presence of blasts in peripheral blood (median: 16%, range: 0 – 99). On average, the LDH level was elevated (median: 450 U/l, range: 102 – 8,840). The parameters of hemostasis showed prolonged PT (70 ± 18%) and very high D-dimer (median: 3.0 μg/ml, range: 0.19 – 138). The criteria for overt disseminated intravascular coagulopathy were met by 74/176 patients (42%), and they had a significantly higher ISTH score (group I – ISTH, median: 5, range: 5 – 7; group II – ISTH, median: 3, range: 0 – 3) (p <0.001).

The greatest number of patients had AML FAB type M4 (n = 66, 37.5%) and belonged to the ELN group of intermediate cytogenetic risk (n = 93, 52.8%). CD56 positivity was established in 62 patients (35.2%). Patients also had significant associated diseases (HCT-CI score, median: 1, range: 0 – 8).

Patients with DIC were significantly older (57.4 ± 12.4 years), as compared to patients without DIC (51.2 ± 15.5 years), p = 0.006. Hemorrhagic syndrome at the time of diagnosis was significantly more common in the group of patients with DIC, 39/74 (52.7%), as compared to patients without DIC, 33/102 (32.4%), p = 0.01. As to laboratory parameters, patients with DIC had a significantly lower platelet count (group I – median: 32.5x10⁹ /l, range: 1 – 151, as compared to group II – median: 61.5, range: 2 – 421; p < 0.001), significantly higher levels of LDH (group I – median: 591.5 U/l, range: 102 – 5,786; group II – median: 383 U/l, range: 108 – 8,840; p < 0.001), significantly longer PT (group I: 61.6±16%; group II: 78±15%; p < 0.001), and significantly higher levels of D-dimer (group I – 6.2 μg/ml, range: 0.82 – 138; group II – 1.32 μg/ml, range: 0.2 – 74; p < 0.001). Patients with DIC had a greater number of comorbidities (group I – average HCT-CI score: 2, range: 0 – 8), as compared to the patients without DIC (group II – average HCT-CI score: 1, range: 0 – 6), p = 0.01. The type of AML, the ELN cytogenetic risk group, and CD56 positivity, did not affect the development of overt disseminated intravascular coagulopathy (p > 0.05).

As to the outcome, by the end of the follow-up period, of the 176 subjects, there were 48 living non-APL AML patients (27.7%). The occurrence of DIC did not affect the outcome (living, group I – 18/74 (24.3%); living, group II – 30/102 (29.4%), p = 0.496). Induction death was registered in 41/176 patients (23.3%), and it did not significantly differ between the two analyzed groups (group I – n = 20/74 (27%); group II – n = 21/102 (27%); p = 0.291).

The overall survival of all patients was 7 months (range: 0 – 57), and though it was shorter in patients with overt disseminated intravascular coagulopathy (group I – 5 months, range: 0 – 57), as compared to patients without coagulopathy (group II – 7 months, range: 0 – 49), this difference did not show statistical significance (log-rank: 0.518), which has been presented with the Kaplan Meier survival curve (Figure 1).

Figure 1. Survival (months) of 176 patients with non-APL AML, in relation to overt DIC

DISCUSSION

The pathophysiological mechanism of DIC development in acute leukemias is complex, and it simultaneously includes the following: a) coagulation activation caused by the exposure of the tissue factor (TF) to blood; b) the disruption of the anticoagulant mechanism control, and c) the suppression of fibrinolysis through increased expression of PAI-1 (plasminogen activator inhibitor-1). These changes jointly cause endothelial disfunction and microvascular thromboses, which lead to organ dysfunction, and have a significantly negative effect on the prognosis of the underlying disease [14]. Leukemia cells release TF, and they also secrete proinflammatory cytokines, primarily IL-6 (interleukin) and TNF alfa (tumor necrosis factor), which damage the endothelium of blood vessels. Endothelial damage, on the one hand, leads to increased expression of TF and PAI-1, and, on the other hand, it causes decreased expression of thrombomodulin (TM), which converts protein C (PC) into activated PC (APC), and inhibits coagulation. Leukemia cells release many microparticles, which, in addition to TF, also contain cancer procoagulant, which has the activity of serine proteas, and which initiates the coagulation cascade as well as the generating of thrombin through direct activation of factor X (FX).

DIC is characterized by fibrinolysis disruption, whose purpose, in physiological conditions, is to prevent insufficiency in peripheral circulation through the degradation of fibrin deposits. In acute leukemias, proinflammatory cytokines cause the increased expression of PAI-1, which, through the inhibition of plasminogen activators, prevents the synthesis of plasmin (secondary fibrinolysis). In acute leukemias, primary fibrinolysis is also very intensified. All of this leads to marked hyperfibrinogenemia and the increase in fibrinogen/fibrin degradation products (FDP) and D-dimer [18]. Around 15% of patients with non-APL AML additionally develop DIC during induction remission therapy. Malignant cells, under the influence of cytotoxic drugs, undergo apoptosis, releasing intranuclear proteins, such as histone H3 and HMGB1 (high-mobility group box-1), which contributes to the development of DIC and the tumor lysis syndrome (TLS) [19].

The incidence of DIC in acute leukemias is variable. The highest incidence is in APL, and the lowest incidence is in B-cell acute lymphoblastic leukemia [16]. Disseminated intravascular coagulopathy has been intensively tested and analyzed in APL, however, the data on its frequency and significance in non-APL AML are very limited and varied. Hence, reported frequency of DIC, based on ISTH criteria, in patients with non-APL AML, at diagnosis, ranges from 6.4% [20] to 25.2% [16]. Libourel et al. [21] reported that the frequency of disseminated intravascular coagulopathy, determined according to the ISTH criteria, was higher in younger patients (18 – 65 years) (8.5%), as compared to older patients with nonAPL AML (6.3%), while older patients significantly more often had hemorrhagic syndrome (13%). The frequency of DIC in the patients from our study was 42%, which is significantly higher, as compared to the abovementioned data. Additionally, our patients more often had manifest bleeding at diagnosis, more than 50% of the patients with DIC (39/74) had hemorrhagic syndrome.

Leukocytosis (>20x10⁹ /l) carries a higher risk for the development of DIC in AML. Leukemia cells, which, conversely to erythrocytes, are not elastic and flexible, create accumulations in the microcirculation, which, on one hand, leads to vascular occlusion and additional damage to the endothelium, and, on the other hand, it leads to increased release of cytokines, microparticles, and intranuclear proteins from aggregated blasts [16],[22]. In addition to marked leukocytosis, in our group of patients with DIC, significantly higher levels of LDH were also registered. A high concentration of LDH is a predictor of a high risk of bleeding [23].

It is important to emphasize that patients with DIC were significantly older and that they had significantly more comorbidities, as compared to patients with non-APL AML who did did not develop DIC. Elderly age, in itself, is considered to be a state of chronic inflammation [24], while present associated diseases can additionally affect the development of DC. We were not able to find analyses of the influence of comorbidities on the development of DIC in available literature.

The patients with DIC, from our study, had a more severe degree of thrombocytopenia, significantly longer PT, and a higher D-dimer, as compared to patients without DIC, and these are the parameters that are monitored on the ISTH DIC score. Establishing the correct diagnosis of DIC in AML is often a challenge, bearing in mind the nature of the disease itself [11]. As patients with leukemia very often suffer from thrombocytopenia, even in the absence of DIC, due to the infiltration of bone marrow and the application of cytotoxic therapy, a new JMWH (Japanese Ministry of Health and Welfare) scoring system has been proposed for DIC diagnosis. This system scores the underlying disease, as well (acute leukemias are scored with one point), however, the scoring criteria for thrombocytopenia, fibrinogen concentration, and FDP have been modified [25].

As far as the biological characteristics of non-APL AML and the risks for the development of DIC are concerned, in our patients, no difference was determined between the FAB subtypes of AML and the cytogenetic group of risks. Additionally, the expression of CD56, which has been reported to carry an unfavorable prognosis in AML [26], did not affect the development of DIC. Our results differ from the results obtained in the study by Guo et al. [16], who found that the prevalence of DIC was significantly higher in patients with a normal karyotype and the mutations NPM1 and/or FLT3-ITD. Libourel et al. [21] reported the highest frequency of DIC in the FAB type M5. It is interesting to note that the mutations NPM1 and FLT3-ITD [27], as well as the AML type FAB5 [28] have been linked to hyperleukocytosis.

The joint occurrence of DIC and AML carries an unfavorable prognosis [11],[15],[18],[19],[20],[21],[22]. The patients with DIC, from our study, received, not only treatment for their underlying disease, but were simultaneously also given supportive blood derivatives and components therapy [11],[13]. The application of antifibrinolytics in the treatment of DIC in AML is not recommended, due to the danger of promoting the creation of fibrin deposits [11]. In Japan, the application of recombinant soluble thrombomodulin (rTM) has been approved, for treating DIC in acute leukemias, as it inactivates coagulation by binding to thrombin [29]. The outcome, early mortality, and survival in our non-APL AML patients with DIC did not significantly differ from the same parameters in our patients without DIC, which is the consequence of the above-described treatment approach.

CONCLUSION

Based on the research conducted within this study, we can conclude that older age, the presence of comorbidities, leukocytosis, and high levels of LDH, carry a significant risk of DIC development in patients with non-APL AML. The occurrence of overt disseminated intravascular coagulopathy does not negatively affect early mortality, the outcome, and overall survival of patients with non-APL AML, if the diagnosis of DIC is established on time, and timely, appropriate and intensive supportive therapy with blood derivatives and components is administered promptly.